Exposures to organic solvents (including benzene), ionizing radiation and low-frequency electromagnetic fields have been suggested as risk factors for leukemias in adults. Exposure to herbicides of the phenoxyacid variety (2,4,5-T; 2,4-D) has been reported to be associated with an increased risk of non-Hodgkin's lymphomas. In addition, several studies have suggested that agricultural workers have an excess risk of hemato- lymphopoietic malignancies, although no specific agent has yet been identified. Italy is particularly attractive for investigating further these associations, because: a) there is a high prevalance of exposure to solvents in industrial areas (estimated to be 10-16% in the provinces of Torino, Varese and Firenze), and a high prevalence of exposure to herbicides (including phenoxyacids) in agricultural areas; b) there is also considerable extra-occupational exposure to solvents and pesticides; c) competent groups of hematologists exist in several towns, using advanced diagnostic techniques including cytogenetics. We propose therefore to conduct a multicenter population-based case-control study in 12 Italian areas, chosen on the basis of a high exposure prevalence. In three years, we intend to interview approximately 1060 cases of leukemia, 1880 cases of lymphoma (Hodgkin and non-Hodgkin) and 400 cases of myeloma; 2000 randomly selected individuals, residents of the same areas, will serve as a population control group. Cases will be all those occurring in the study period among residents of the 12 areas, newly diagnosed. Particular care will be dedicated to diagnostic aspects, including B- and T-cell characterization and, in four of the areas, cytogenetics using banding techniques. Cases and controls will be blindly interviewed by trained personnel. The occupational histories will be blindly evaluated by industrial hygienists and agronomists, in order to assess the probability of exposure to organic solvents, phenoxy herbicides and other chemicals. The study will have a statistical power of 90% to detect a relative risk of 2.0 for the associations between organic solvent exposure and leukemias and between herbicide exposure and non-Hodgkin's lymphomas in men (p=0.05).